The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses subprocess.run to execute git commands for repository metadata and history retrieval (e.g., git log, git diff, git status in external_context.py and git_index.py). It also executes node to run a local JavaScript/TypeScript analysis bridge script (ts_ast_bridge.js) included with the skill.
[DYNAMIC_EXECUTION]: The ts_ast_bridge.js script dynamically resolves and requires the typescript package from the target project's node_modules or local environment to perform AST-based analysis.
[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted code and data from the target repository, creating a potential vector for indirect prompt injection if malicious instructions are embedded in the source files.
Ingestion points: generate.py reads file content from the project directory; git_index.py reads git log and blame data.
Boundary markers: Output is structured as JSON; SKILL.md defines specific contracts for how the agent should interpret read and report payloads.
Capability inventory: Subprocess execution (git, node), file system read access, and local file system write access to the {project}/repo/progress/ directory.
Sanitization: Uses json.dumps for serialization and regex-based chunking to delimit content windows.

context-codebase